The Hatten Lab Projects


3D Imaging




The ability to image living cells provides an exciting opportunity to monitor both population dynamics as well as the behavior of individual cells. The goal of this project is to use newly developed compensation algorithms to render the detailed morphologies of fluorescent cells (e.g. Astn-GFP tagged cells) in three dimensions, as they move through developing brain tissue. For these experiments we will image GFP-labeled cells in both whole mount preparations (embryonic cortex) and thick vibratome sections (postnatal cerebellum) prepared from Astn-GFP animals. The expression of Astn by LIVING CELLS can be monitored by video microscopy of cells in tissue slices of embryonic and early postnatal cerebral cortex and cerebellar cortex. Cells within brain tissue can be imaged with microscopes fitted with either a confocal laser scan head or with a video camera that has a cooled-charged coupled device (cooled CCD). In either case, an automatic focusing system (Ludl drive), controlled by the host PC hardware and software, provides automated stepwise focus through the Z axis.
The microscope facility is a suite of acquisition stations (Wintel PC's) and imaging/animation stations (SGI O2). Images and animations can be distributed via the Internet in MPEG-1 format (created with MetaMovie software), or recorded to S-VHS videotape directly from the O2. The heterogeneous nature of this suite provided a compelling reason to develop all in-house software in the Java programming language, due to Java's cross platform capabilities.

3d imaging by Peter Selmayr
Cell Migration MPEG movie

(MPEG movie, 4.6MB) Hatten, Bhatt, Tomoda, Didkovsky, 2000

 Migrating Granule Cell in vivo

With this system, we can image living cells in tissue slices, ship the images to the SGI workstation, use Tween Machine to choose a set of keyframes and execute the commands for Voxel View to image the cells within our sample. We can then use Voxel Animator to create movies of labeled cells or individual cells within these fields. We use commercial software as well as software developed in-house to provide compensation algorithms that allow us to render 3D animation of complex, contoured surfaces like a neuron and its processes. These programs provide enhanced masking capabilities to digitally isolate ("seed") a particular cell or axon. This feature allows us to view the form and behavior of individual cells within the large set labeled cells in thick slabs of tissue. With the microscope facilities and computer-based image analysis programs we can acquire detailed views of labeled cells in situ.


Cell Migration MPEG movie

(MPEG movie, 5.5MB)
Hatten, Bhatt, Tomoda, Didkovsky, 2000

 Migrating Granule Cell in vivo


Completed aspects of these experiments include the development of Java software that facilitates the transfer of files from the microscope set-ups to the Voxel View Computer System (MetaVox), animation preview/script generating software (TweenMachine), and software to perform 3D Volume voxel processing operations (VoxelMincer).

Ongoing aspects of this work includes the development of training media to aid in the visualization of 3D microscopy and volume manipulation. The movies below demonstrate how the illusion of a 3D solid is created by stacking 2D planes, how that data is transferred from the PC's running MetaMorph to the SGI running VoxelView, and finally, how the virtual camera in VoxelView is oriented around a 3D volume.

Training Movie #1: A 3D volume is a stack of 2D planes.
3d imaging by Peter Selmayr 3d imaging by Peter Selmayr 3d imaging by Peter Selmayr


Training Movie #2: 3D file conversion and transport.
3d imaging by Peter Selmayr 3d imaging by Peter Selmayr 3d imaging by Peter Selmayr


Training Movie #3: Camera orientation in 3D space.
3d imaging by Peter Selmayr 3d imaging by Peter Selmayr 3d imaging by Peter Selmayr



Work is in progress to use emerging versions of JAVA combining software-generated models of neurons with real image data in a virtual environment which can be navigated and explored in real time. A further goal will be to write JAVA-based programs to speed up the process of generating time-lapse movies in thick tissue slices. In these experiments, we will be acquiring large stacks of 100-200 images, each comprising approximately 1µm in depth.
Hatten Lab Projects
Hatten Lab Home Page